diff --git a/.gitattributes b/.gitattributes index a6344aac8c09253b3b630fb776ae94478aa0275b..a09db2ea4d1bd1fc1c09f6fca45e8e4054953535 100644 --- a/.gitattributes +++ b/.gitattributes @@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text *.zip filter=lfs diff=lfs merge=lfs -text *.zst filter=lfs diff=lfs merge=lfs -text *tfevents* filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text +model.safetensors.index.json filter=lfs diff=lfs merge=lfs -text diff --git a/.quant_summary.txt b/.quant_summary.txt new file mode 100644 index 0000000000000000000000000000000000000000..bc471cf0028fd8ccdc249da3b6da18291ed68da5 --- /dev/null +++ b/.quant_summary.txt @@ -0,0 +1,41596 @@ +model.layers.0.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.0.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.0.mlp.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.1.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.1.mlp.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.2.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.2.mlp.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.3.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.75e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e+00 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e+00 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.11e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.3.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.3.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.3.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.4.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e+00 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.4.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.4.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.4.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.5.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.96e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e+00 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.72e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.41e-01 calibrator=MaxCalibrator quant) +model.layers.5.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.5.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.6.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.82e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e+00 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.6.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.6.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.7.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.99e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.54e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e+00 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e+00 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.71e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.98e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.77e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.17e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.7.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.7.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.8.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.38e+00 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.95e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.95e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e+01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.73e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.8.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.8.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.8.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.9.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e+00 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.56e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e+01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.9.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.9.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.9.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.10.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.33e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e+00 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.69e+00 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.10.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.10.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.10.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.11.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.64e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.59e+00 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.11.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.11.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.12.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.84e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e+00 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.12.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.12.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.13.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.36e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e+00 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.13.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.13.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.14.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.71e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.69e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e+00 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.14.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.14.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.15.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.81e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e+00 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.98e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.15.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.15.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.16.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e+00 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.16.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.16.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.17.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.88e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e+00 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.17.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.17.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.17.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.18.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e+00 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.88e+00 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.18.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.18.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.19.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e+00 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.06e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.19.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.19.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.19.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.20.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e+00 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.86e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.86e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.67e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e+01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.20.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.20.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.20.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.21.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e+00 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e+00 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.21.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.21.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.21.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.22.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e+00 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.82e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e+01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.22.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.22.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.22.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.23.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e+00 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.38e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e+00 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.23.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.23.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.23.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.24.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e+00 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.12e+00 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.24.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.24.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.25.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e+00 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e+00 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.25.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.25.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.25.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.26.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e+00 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e+00 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.26.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.26.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.27.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e+00 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e+00 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.27.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.27.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.27.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.28.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e+00 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.55e+01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.28.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.28.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.28.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.29.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e+00 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e+01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.29.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.29.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.29.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.30.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e+00 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.58e+01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.30.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.30.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.31.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e+00 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.66e+00 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.23e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.31.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.31.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.31.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.32.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e+00 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.74e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e+00 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.04e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.32.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.32.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.32.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.33.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e+00 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.25e+00 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.33.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.33.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.34.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e+00 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.52e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.28e+00 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.34.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.34.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.34.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.35.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e+00 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.31e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e+00 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.35.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.35.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.35.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.36.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e+00 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.59e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.26e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.06e+00 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.36.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.36.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.36.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.37.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e+00 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.86e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.77e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.86e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.88e+00 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.80e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.37.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.37.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.38.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e+00 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.98e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.65e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.69e+00 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.79e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.38.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.38.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.38.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.39.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e+00 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.91e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.00e+00 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.39.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.39.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.39.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.40.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e+00 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.31e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.61e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.12e+00 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.40.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.40.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.41.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e+00 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e+01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.41.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.41.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.41.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.42.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e+00 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.84e+00 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.42.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.42.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.42.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.43.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e+00 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.41e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.38e+00 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.43.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.43.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.43.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.44.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e+00 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.55e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.85e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.36e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.88e+00 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.44.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.44.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.45.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e+00 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.38e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.47e+00 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.45.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.45.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.46.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e+00 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.25e+00 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.46.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.46.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.46.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.47.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e+00 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e+01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.15e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.25e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.47.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.47.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.48.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e+00 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.84e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e+01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.48.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.48.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.49.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e+00 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.50e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e+01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.33e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.49.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.49.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.50.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e+00 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e+01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.50.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.50.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.51.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e+00 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e+01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.51.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.51.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.51.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.52.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.59e+00 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e+01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.52.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.52.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.53.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e+00 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e+01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.53.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.53.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.53.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.54.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.92e+00 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.44e+00 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.54.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.54.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.54.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.55.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.00e+00 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.91e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e+01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.55.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.55.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.56.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e+00 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.31e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e+01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.56.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.56.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.57.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e+00 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.94e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e+01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.57.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.57.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.58.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.08e+00 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.69e+00 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.58.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.58.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.59.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e+00 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.22e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.44e+00 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.20e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.59.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.59.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.59.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.60.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e+00 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e+01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.60.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.60.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.60.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.61.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.17e+00 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.77e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.31e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.70e+01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.43e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.45e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.61.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.61.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.62.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e+00 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.76e+01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.30e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.62.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.62.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.63.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e+00 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.69e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e+00 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e+00 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.63.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.63.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.64.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.94e+00 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.45e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.98e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e+01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.64.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.64.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.65.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.06e+00 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.92e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.84e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.25e+00 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.54e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.65.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.65.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.65.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.66.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e+00 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.58e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.47e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.61e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e+01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.66.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.66.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.67.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e+00 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.65e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.36e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.78e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.96e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e+01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.35e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.43e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.50e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.67.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.67.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.68.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.41e+00 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e+00 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e+00 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.67e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.98e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.28e+01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.69e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.64e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.97e-01 calibrator=MaxCalibrator quant) +model.layers.68.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.68.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.68.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.69.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.08e+00 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.68e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.46e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.68e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.73e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.70e+01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.69.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.69.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.69.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.70.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e+00 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.90e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.65e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.70e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.50e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.83e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.37e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e+00 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e+00 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e+02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.89e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.74e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.59e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.70.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.70.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.70.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.71.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.06e+00 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.29e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e+00 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.54e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.68e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.95e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.50e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e+01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.22e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.71.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.71.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.72.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.19e+00 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e+00 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.56e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.80e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.24e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e+01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.46e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.84e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.72.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.72.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.72.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.73.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.59e+00 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.67e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.85e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e+01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.57e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.59e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.63e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.84e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.81e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.94e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.83e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.73.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.73.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.73.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.74.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.41e+00 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.62e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.65e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.96e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.81e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.05e+01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.03e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.08e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.89e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.37e-01 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.28e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.74.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.74.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.74.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.75.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e+00 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.90e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.95e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.39e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e+02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.77e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.13e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.39e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.52e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.47e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.77e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.77e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.73e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.36e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.42e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.71e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.79e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.86e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.77e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.23e-02 calibrator=MaxCalibrator quant) +model.layers.75.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.75.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.75.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.76.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e+00 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.68e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.70e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.39e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.90e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.90e+01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.30e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.44e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.96e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.79e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.33e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.98e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.01e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.06e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.50e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.55e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.00e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.38e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.91e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.13e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.24e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.62e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.67e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.81e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.72e-02 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.76.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.76.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_a_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_a_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_a_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_a_proj_with_mqa.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_a_proj_with_mqa.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_a_proj_with_mqa.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_b_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_b_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.kv_b_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.o_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.o_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.o_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.wq_b.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.wq_b.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.wq_b.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.wk.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.wk.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.wk.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.k_norm.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.k_norm.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.weights_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.weights_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.indexer.weights_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.q_bmm_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.k_bmm_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.v_bmm_quantizer TensorQuantizer(disabled) +model.layers.77.self_attn.softmax_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.experts.gate_up_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.61e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.69e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.80e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.73e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.98e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.68e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.06e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.56e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.17e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.47e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.72e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.44e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.13e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.02e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.22e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.50e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.97e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.10e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.33e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.85e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.74e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.91e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.60e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.gate_up_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_input_quantizer TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e+02 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.0 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.1 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.17e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.2 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.17e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.3 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.86e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.4 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.5 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.6 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.7 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.31e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.8 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.9 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.10 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.11 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.12 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.13 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.14 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.30e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.15 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.16 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.17 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.64e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.18 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.19e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.19 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.20 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.31e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.21 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.27e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.22 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.23 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.24 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.28e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.25 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.26 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.27 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.02e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.28 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.29 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.69e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.30 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.31 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.32 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.33 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.39e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.34 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.35 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.36 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.77e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.37 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.38 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.45e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.39 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.40 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.41 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.42 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.43 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.44 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.81e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.45 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.46 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.47 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.48 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.53e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.49 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.50 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.51 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.52 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.53 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.54 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.55 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.18e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.56 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.99e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.57 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.22e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.58 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.59 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.49e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.60 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.61 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.24e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.62 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.74e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.63 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.64 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.25e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.65 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.53e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.66 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.37e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.67 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.68 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.69 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.70 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.71 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.72 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.73 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.42e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.74 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.75 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.76 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.77 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.67e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.78 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.18e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.79 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.80 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.68e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.81 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.82 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.25e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.83 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.84 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.85 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.86 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.87 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.00e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.88 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.37e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.89 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.81e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.90 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.15e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.91 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.92 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.80e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.93 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.94 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.57e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.95 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.92e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.96 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.97 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.98 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.99 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.100 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.21e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.101 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.05e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.102 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.103 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.34e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.104 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.105 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.94e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.106 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.107 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.87e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.108 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.63e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.109 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.04e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.110 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.97e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.111 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.15e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.112 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.113 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.114 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.51e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.115 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.11e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.116 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.117 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.118 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.119 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.03e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.120 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.54e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.121 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.122 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.123 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.124 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.125 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.26e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.126 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.05e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.127 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.00e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.128 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.71e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.129 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.00e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.130 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.61e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.131 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.62e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.132 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.133 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.134 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.135 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.136 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.137 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.04e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.138 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.139 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.140 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.98e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.141 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.62e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.142 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.03e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.143 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.144 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.145 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.146 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.147 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.148 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.149 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.82e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.150 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.38e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.151 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.73e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.152 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.58e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.153 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.94e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.154 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.36e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.155 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.43e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.156 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.23e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.157 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.158 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.01e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.159 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.160 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.161 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.86e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.162 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.163 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.164 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.165 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.166 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.12e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.167 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.30e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.168 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.169 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.170 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.46e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.171 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.172 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.35e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.173 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.174 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.175 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.29e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.176 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.177 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.96e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.178 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.179 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.51e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.180 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.92e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.181 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.32e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.182 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.183 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.57e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.184 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.185 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.27e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.186 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.187 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.24e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.188 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.40e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.189 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.02e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.190 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.191 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.88e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.192 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.20e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.193 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.20e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.194 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.70e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.195 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.93e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.196 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.197 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.59e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.198 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.199 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.200 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.201 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.70e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.202 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.203 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.84e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.204 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.205 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.66e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.206 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.12e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.207 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.44e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.208 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.209 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.210 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.59e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.211 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.07e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.212 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.76e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.213 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.61e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.214 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.28e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.215 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.216 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.217 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.218 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=5.16e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.219 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=7.89e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.220 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.221 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.65e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.222 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.223 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.95e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.224 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.14e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.225 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.60e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.226 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.09e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.227 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.34e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.228 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.229 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.48e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.230 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.231 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.63e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.232 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.53e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.233 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.78e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.234 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.83e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.235 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=6.13e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.236 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.237 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.75e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.238 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=4.08e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.239 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.240 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.04e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.241 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.55e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.242 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=dynamic calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.243 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.25e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.244 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.15e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.245 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.41e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.246 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=9.57e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.247 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=8.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.248 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.42e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.249 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.03e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.250 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.52e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.251 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.09e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.252 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=3.79e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.253 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.95e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.254 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=2.11e-01 calibrator=MaxCalibrator quant) +model.layers.77.mlp.experts.down_proj_weight_quantizers.255 TensorQuantizer((2, 1) bit fake block_sizes={-1: 16, 'type': 'dynamic', 'scale_bits': (4, 3)}, amax=1.08e+00 calibrator=MaxCalibrator quant) +model.layers.77.mlp.shared_experts.gate_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.gate_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.gate_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.up_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.up_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.up_proj.weight_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.down_proj.input_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.down_proj.output_quantizer TensorQuantizer(disabled) +model.layers.77.mlp.shared_experts.down_proj.weight_quantizer TensorQuantizer(disabled) +lm_head.input_quantizer TensorQuantizer(disabled) +lm_head.output_quantizer TensorQuantizer(disabled) +lm_head.weight_quantizer TensorQuantizer(disabled) +41595 TensorQuantizers found in model diff --git a/chat_template.jinja b/chat_template.jinja new file mode 100644 index 0000000000000000000000000000000000000000..e230020b321d78d161a5cfa505ace4f945e299d5 --- /dev/null +++ b/chat_template.jinja @@ -0,0 +1,119 @@ +[gMASK] +{%- if tools -%} +{%- macro tool_to_json(tool) -%} + {%- set ns_tool = namespace(first=true) -%} + {{ '{' -}} + {%- for k, v in tool.items() -%} + {%- if k != 'defer_loading' and k != 'strict' -%} + {%- if not ns_tool.first -%}{{- ', ' -}}{%- endif -%} + {%- set ns_tool.first = false -%} + "{{ k }}": {{ v | tojson(ensure_ascii=False) }} + {%- endif -%} + {%- endfor -%} + {{- '}' -}} +{%- endmacro -%} +<|system|> +# Tools + +You may call one or more functions to assist with the user query. + +You are provided with function signatures within XML tags: + +{% for tool in tools %} +{%- if 'function' in tool -%} + {%- set tool = tool['function'] -%} +{%- endif -%} +{% if tool.defer_loading is not defined or not tool.defer_loading %} +{{ tool_to_json(tool) }} +{% endif %} +{% endfor %} + + +For each function call, output the function name and arguments within the following XML format: +{function-name}{arg-key-1}{arg-value-1}{arg-key-2}{arg-value-2}...{%- endif -%} +{%- macro visible_text(content) -%} + {%- if content is string -%} + {{- content }} + {%- elif content is iterable and content is not mapping -%} + {%- for item in content -%} + {%- if item is mapping and item.type == 'text' -%} + {{- item.text }} + {%- elif item is string -%} + {{- item }} + {%- endif -%} + {%- endfor -%} + {%- else -%} + {{- content }} + {%- endif -%} +{%- endmacro -%} +{%- set ns = namespace(last_user_index=-1, thinking_indices='') -%} +{%- for m in messages %} + {%- if m.role == 'user' %} + {%- set ns.last_user_index = loop.index0 -%} + {%- elif m.role == 'assistant' %} + {%- if m.reasoning_content is string %} + {%- set ns.thinking_indices = ns.thinking_indices ~ ',' ~ ns.last_user_index ~ ',' -%} + {%- endif %} + {%- endif %} +{%- endfor %} +{%- set ns.has_thinking = false -%} +{%- for m in messages -%} +{%- if m.role == 'user' -%}<|user|>{{ visible_text(m.content) }}{% set ns.has_thinking = (',' ~ loop.index0 ~ ',') in ns.thinking_indices -%} +{%- elif m.role == 'assistant' -%} +<|assistant|> +{%- set content = visible_text(m.content) %} +{%- if m.reasoning_content is string %} + {%- set reasoning_content = m.reasoning_content %} +{%- elif '' in content %} + {%- set reasoning_content = content.split('')[0].split('')[-1] %} + {%- set content = content.split('')[-1] %} +{%- elif loop.index0 > ns.last_user_index and not (enable_thinking is defined and not enable_thinking) %} + {%- set reasoning_content = '' %} +{%- elif loop.index0 < ns.last_user_index and ns.has_thinking %} + {%- set reasoning_content = '' %} +{%- endif %} +{%- if ((clear_thinking is defined and not clear_thinking) or loop.index0 > ns.last_user_index) and reasoning_content is defined -%} +{{ '' + reasoning_content + ''}} +{%- else -%} +{{ '' }} +{%- endif -%} +{%- if content.strip() -%} +{{ content.strip() }} +{%- endif -%} +{% if m.tool_calls %} +{% for tc in m.tool_calls %} +{%- if tc.function %} + {%- set tc = tc.function %} +{%- endif %} +{{- '' + tc.name -}} +{% set _args = tc.arguments %}{% for k, v in _args.items() %}{{ k }}{{ v | tojson(ensure_ascii=False) if v is not string else v }}{% endfor %}{% endfor %} +{% endif %} +{%- elif m.role == 'tool' -%} +{%- if loop.first or (messages[loop.index0 - 1].role != "tool") %} + {{- '<|observation|>' -}} +{%- endif %} +{%- if m.content is string -%} + {{- '' + m.content + '' -}} +{%- elif m.content is iterable and m.content is not mapping and m.content and m.content.0.type == "tool_reference" -%} + {{- '\n' -}} + {% for tr in m.content %} + {%- for tool in tools -%} + {%- if 'function' in tool -%} + {%- set tool = tool['function'] -%} + {%- endif -%} + {%- if tool.name == tr.name -%} + {{- tool_to_json(tool) + '\n' -}} + {%- endif -%} + {%- endfor -%} + {%- endfor -%} + {{- '' -}} +{%- else -%} + {{- '' + visible_text(m.content) + '' -}} +{% endif -%} +{%- elif m.role == 'system' -%} +<|system|>{{ visible_text(m.content) }} +{%- endif -%} +{%- endfor -%} +{%- if add_generation_prompt -%} + <|assistant|>{{- '' if (enable_thinking is defined and not enable_thinking) else '' -}} +{%- endif -%} \ No newline at end of file diff --git a/config.json b/config.json new file mode 100644 index 0000000000000000000000000000000000000000..184d3c402fbdc14178b0e07ca0d3534e3906effc --- /dev/null +++ b/config.json @@ -0,0 +1,402 @@ +{ + "architectures": [ + "GlmMoeDsaForCausalLM" + ], + "attention_bias": false, + "attention_dropout": 0.0, + "bos_token_id": 0, + "dtype": "bfloat16", + "eos_token_id": [ + 154820, + 154827, + 154829 + ], + "ep_size": 1, + "first_k_dense_replace": 3, + "hidden_act": "silu", + "hidden_size": 6144, + "index_head_dim": 128, + "index_n_heads": 32, + "index_topk": 2048, + "indexer_rope_interleave": true, + "indexer_types": [ + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full", + "full" + ], + "initializer_range": 0.02, + "intermediate_size": 12288, + "kv_lora_rank": 512, + "max_position_embeddings": 202752, + "mlp_layer_types": [ + "dense", + "dense", + "dense", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse", + "sparse" + ], + "model_type": "glm_moe_dsa", + "moe_intermediate_size": 2048, + "moe_layer_freq": 1, + "n_group": 1, + "n_routed_experts": 256, + "n_shared_experts": 1, + "norm_topk_prob": true, + "num_attention_heads": 64, + "num_experts_per_tok": 8, + "num_hidden_layers": 78, + "num_key_value_heads": 64, + "num_nextn_predict_layers": 1, + "pad_token_id": 154820, + "pretraining_tp": 1, + "q_lora_rank": 2048, + "qk_head_dim": 256, + "qk_nope_head_dim": 192, + "qk_rope_head_dim": 64, + "rms_norm_eps": 1e-05, + "rope_interleave": true, + "rope_parameters": { + "rope_theta": 1000000, + "rope_type": "default" + }, + "routed_scaling_factor": 2.5, + "scoring_func": "sigmoid", + "tie_word_embeddings": false, + "topk_group": 1, + "topk_method": "noaux_tc", + "transformers_version": "5.8.0", + "use_cache": true, + "v_head_dim": 256, + "vocab_size": 154880, + "quantization_config": { + "config_groups": { + "group_0": { + "input_activations": { + "dynamic": false, + "num_bits": 4, + "type": "float", + "group_size": 16 + }, + "weights": { + "dynamic": false, + "num_bits": 4, + "type": "float", + "group_size": 16 + }, + "targets": [ + "Linear" + ] + } + }, + "ignore": [ + "lm_head", + "model.layers.0*", + "model.layers.1.*", + "model.layers.10.mlp.shared_experts*", + "model.layers.10.self_attn*", + "model.layers.11.mlp.shared_experts*", + "model.layers.11.self_attn*", + "model.layers.12.mlp.shared_experts*", + "model.layers.12.self_attn*", + "model.layers.13.mlp.shared_experts*", + "model.layers.13.self_attn*", + "model.layers.14.mlp.shared_experts*", + "model.layers.14.self_attn*", + "model.layers.15.mlp.shared_experts*", + "model.layers.15.self_attn*", + "model.layers.16.mlp.shared_experts*", + "model.layers.16.self_attn*", + "model.layers.17.mlp.shared_experts*", + "model.layers.17.self_attn*", + "model.layers.18.mlp.shared_experts*", + "model.layers.18.self_attn*", + "model.layers.19.mlp.shared_experts*", + "model.layers.19.self_attn*", + "model.layers.2.*", + "model.layers.20.mlp.shared_experts*", + "model.layers.20.self_attn*", + "model.layers.21.mlp.shared_experts*", + "model.layers.21.self_attn*", + "model.layers.22.mlp.shared_experts*", + "model.layers.22.self_attn*", + "model.layers.23.mlp.shared_experts*", + "model.layers.23.self_attn*", + "model.layers.24.mlp.shared_experts*", + "model.layers.24.self_attn*", + "model.layers.25.mlp.shared_experts*", + "model.layers.25.self_attn*", + "model.layers.26.mlp.shared_experts*", + "model.layers.26.self_attn*", + "model.layers.27.mlp.shared_experts*", + "model.layers.27.self_attn*", + "model.layers.28.mlp.shared_experts*", + "model.layers.28.self_attn*", + "model.layers.29.mlp.shared_experts*", + "model.layers.29.self_attn*", + "model.layers.3.mlp.shared_experts*", + "model.layers.3.self_attn*", + "model.layers.30.mlp.shared_experts*", + "model.layers.30.self_attn*", + "model.layers.31.mlp.shared_experts*", + "model.layers.31.self_attn*", + "model.layers.32.mlp.shared_experts*", + "model.layers.32.self_attn*", + "model.layers.33.mlp.shared_experts*", + "model.layers.33.self_attn*", + "model.layers.34.mlp.shared_experts*", + "model.layers.34.self_attn*", + "model.layers.35.mlp.shared_experts*", + "model.layers.35.self_attn*", + "model.layers.36.mlp.shared_experts*", + "model.layers.36.self_attn*", + "model.layers.37.mlp.shared_experts*", + "model.layers.37.self_attn*", + "model.layers.38.mlp.shared_experts*", + "model.layers.38.self_attn*", + "model.layers.39.mlp.shared_experts*", + "model.layers.39.self_attn*", + "model.layers.4.mlp.shared_experts*", + "model.layers.4.self_attn*", + "model.layers.40.mlp.shared_experts*", + "model.layers.40.self_attn*", + "model.layers.41.mlp.shared_experts*", + "model.layers.41.self_attn*", + "model.layers.42.mlp.shared_experts*", + "model.layers.42.self_attn*", + "model.layers.43.mlp.shared_experts*", + "model.layers.43.self_attn*", + "model.layers.44.mlp.shared_experts*", + "model.layers.44.self_attn*", + "model.layers.45.mlp.shared_experts*", + "model.layers.45.self_attn*", + "model.layers.46.mlp.shared_experts*", + "model.layers.46.self_attn*", + "model.layers.47.mlp.shared_experts*", + "model.layers.47.self_attn*", + "model.layers.48.mlp.shared_experts*", + "model.layers.48.self_attn*", + "model.layers.49.mlp.shared_experts*", + "model.layers.49.self_attn*", + "model.layers.5.mlp.shared_experts*", + "model.layers.5.self_attn*", + "model.layers.50.mlp.shared_experts*", + "model.layers.50.self_attn*", + "model.layers.51.mlp.shared_experts*", + "model.layers.51.self_attn*", + "model.layers.52.mlp.shared_experts*", + "model.layers.52.self_attn*", + "model.layers.53.mlp.shared_experts*", + "model.layers.53.self_attn*", + "model.layers.54.mlp.shared_experts*", + "model.layers.54.self_attn*", + "model.layers.55.mlp.shared_experts*", + "model.layers.55.self_attn*", + "model.layers.56.mlp.shared_experts*", + "model.layers.56.self_attn*", + "model.layers.57.mlp.shared_experts*", + "model.layers.57.self_attn*", + "model.layers.58.mlp.shared_experts*", + "model.layers.58.self_attn*", + "model.layers.59.mlp.shared_experts*", + "model.layers.59.self_attn*", + "model.layers.6.mlp.shared_experts*", + "model.layers.6.self_attn*", + "model.layers.60.mlp.shared_experts*", + "model.layers.60.self_attn*", + "model.layers.61.mlp.shared_experts*", + "model.layers.61.self_attn*", + "model.layers.62.mlp.shared_experts*", + "model.layers.62.self_attn*", + "model.layers.63.mlp.shared_experts*", + "model.layers.63.self_attn*", + "model.layers.64.mlp.shared_experts*", + "model.layers.64.self_attn*", + "model.layers.65.mlp.shared_experts*", + "model.layers.65.self_attn*", + "model.layers.66.mlp.shared_experts*", + "model.layers.66.self_attn*", + "model.layers.67.mlp.shared_experts*", + "model.layers.67.self_attn*", + "model.layers.68.mlp.shared_experts*", + "model.layers.68.self_attn*", + "model.layers.69.mlp.shared_experts*", + "model.layers.69.self_attn*", + "model.layers.7.mlp.shared_experts*", + "model.layers.7.self_attn*", + "model.layers.70.mlp.shared_experts*", + "model.layers.70.self_attn*", + "model.layers.71.mlp.shared_experts*", + "model.layers.71.self_attn*", + "model.layers.72.mlp.shared_experts*", + "model.layers.72.self_attn*", + "model.layers.73.mlp.shared_experts*", + "model.layers.73.self_attn*", + "model.layers.74.mlp.shared_experts*", + "model.layers.74.self_attn*", + "model.layers.75.mlp.shared_experts*", + "model.layers.75.self_attn*", + "model.layers.76.mlp.shared_experts*", + "model.layers.76.self_attn*", + "model.layers.77.mlp.shared_experts*", + "model.layers.77.self_attn*", + "model.layers.8.mlp.shared_experts*", + "model.layers.8.self_attn*", + "model.layers.9.mlp.shared_experts*", + "model.layers.9.self_attn*" + ], + "quant_algo": "NVFP4", + "producer": { + "name": "modelopt", + "version": "0.45.0.dev44+gc273ddb8a.d20260509" + }, + "quant_method": "modelopt" + } +} \ No newline at end of file diff --git a/generation_config.json b/generation_config.json new file mode 100644 index 0000000000000000000000000000000000000000..60b737aade6ff85f4665836b3801164a716fb39c --- /dev/null +++ b/generation_config.json @@ -0,0 +1,13 @@ +{ + "_from_model_config": true, + "do_sample": true, + "eos_token_id": [ + 154820, + 154827, + 154829 + ], + "pad_token_id": 154820, + "temperature": 1.0, + "top_p": 0.95, + "transformers_version": "5.8.0" +} diff --git a/hf_quant_config.json b/hf_quant_config.json new file mode 100644 index 0000000000000000000000000000000000000000..ddea5885a147f61f5723580e1883f1975a436bc5 --- /dev/null +++ b/hf_quant_config.json @@ -0,0 +1,167 @@ +{ + "producer": { + "name": "modelopt", + "version": "0.45.0.dev44+gc273ddb8a.d20260509" + }, + "quantization": { + "quant_algo": "NVFP4", + "kv_cache_quant_algo": null, + "group_size": 16, + "exclude_modules": [ + "lm_head", + "model.layers.0*", + "model.layers.1.*", + "model.layers.10.mlp.shared_experts*", + "model.layers.10.self_attn*", + "model.layers.11.mlp.shared_experts*", + "model.layers.11.self_attn*", + "model.layers.12.mlp.shared_experts*", + "model.layers.12.self_attn*", + "model.layers.13.mlp.shared_experts*", + "model.layers.13.self_attn*", + "model.layers.14.mlp.shared_experts*", + "model.layers.14.self_attn*", + "model.layers.15.mlp.shared_experts*", + "model.layers.15.self_attn*", + "model.layers.16.mlp.shared_experts*", + "model.layers.16.self_attn*", + "model.layers.17.mlp.shared_experts*", + "model.layers.17.self_attn*", + "model.layers.18.mlp.shared_experts*", + "model.layers.18.self_attn*", + "model.layers.19.mlp.shared_experts*", + "model.layers.19.self_attn*", + "model.layers.2.*", + "model.layers.20.mlp.shared_experts*", + "model.layers.20.self_attn*", + "model.layers.21.mlp.shared_experts*", + "model.layers.21.self_attn*", + "model.layers.22.mlp.shared_experts*", + "model.layers.22.self_attn*", + "model.layers.23.mlp.shared_experts*", + "model.layers.23.self_attn*", + "model.layers.24.mlp.shared_experts*", + "model.layers.24.self_attn*", + "model.layers.25.mlp.shared_experts*", + "model.layers.25.self_attn*", + "model.layers.26.mlp.shared_experts*", + "model.layers.26.self_attn*", + "model.layers.27.mlp.shared_experts*", + "model.layers.27.self_attn*", + "model.layers.28.mlp.shared_experts*", + "model.layers.28.self_attn*", + "model.layers.29.mlp.shared_experts*", + "model.layers.29.self_attn*", + "model.layers.3.mlp.shared_experts*", + "model.layers.3.self_attn*", + "model.layers.30.mlp.shared_experts*", + "model.layers.30.self_attn*", + "model.layers.31.mlp.shared_experts*", + "model.layers.31.self_attn*", + "model.layers.32.mlp.shared_experts*", + "model.layers.32.self_attn*", + "model.layers.33.mlp.shared_experts*", + "model.layers.33.self_attn*", + "model.layers.34.mlp.shared_experts*", + "model.layers.34.self_attn*", + "model.layers.35.mlp.shared_experts*", + "model.layers.35.self_attn*", + "model.layers.36.mlp.shared_experts*", + "model.layers.36.self_attn*", + "model.layers.37.mlp.shared_experts*", + "model.layers.37.self_attn*", + "model.layers.38.mlp.shared_experts*", + "model.layers.38.self_attn*", + "model.layers.39.mlp.shared_experts*", + "model.layers.39.self_attn*", + "model.layers.4.mlp.shared_experts*", + "model.layers.4.self_attn*", + "model.layers.40.mlp.shared_experts*", + "model.layers.40.self_attn*", + "model.layers.41.mlp.shared_experts*", + "model.layers.41.self_attn*", + "model.layers.42.mlp.shared_experts*", + "model.layers.42.self_attn*", + "model.layers.43.mlp.shared_experts*", + "model.layers.43.self_attn*", + "model.layers.44.mlp.shared_experts*", + "model.layers.44.self_attn*", + "model.layers.45.mlp.shared_experts*", + "model.layers.45.self_attn*", + "model.layers.46.mlp.shared_experts*", + "model.layers.46.self_attn*", + "model.layers.47.mlp.shared_experts*", + "model.layers.47.self_attn*", + "model.layers.48.mlp.shared_experts*", + "model.layers.48.self_attn*", + "model.layers.49.mlp.shared_experts*", + "model.layers.49.self_attn*", + "model.layers.5.mlp.shared_experts*", + "model.layers.5.self_attn*", + "model.layers.50.mlp.shared_experts*", + "model.layers.50.self_attn*", + "model.layers.51.mlp.shared_experts*", + "model.layers.51.self_attn*", + "model.layers.52.mlp.shared_experts*", + "model.layers.52.self_attn*", + "model.layers.53.mlp.shared_experts*", + "model.layers.53.self_attn*", + "model.layers.54.mlp.shared_experts*", + "model.layers.54.self_attn*", + "model.layers.55.mlp.shared_experts*", + "model.layers.55.self_attn*", + "model.layers.56.mlp.shared_experts*", + "model.layers.56.self_attn*", + "model.layers.57.mlp.shared_experts*", + "model.layers.57.self_attn*", + "model.layers.58.mlp.shared_experts*", + "model.layers.58.self_attn*", + "model.layers.59.mlp.shared_experts*", + "model.layers.59.self_attn*", + "model.layers.6.mlp.shared_experts*", + "model.layers.6.self_attn*", + "model.layers.60.mlp.shared_experts*", + "model.layers.60.self_attn*", + "model.layers.61.mlp.shared_experts*", + "model.layers.61.self_attn*", + "model.layers.62.mlp.shared_experts*", + "model.layers.62.self_attn*", + "model.layers.63.mlp.shared_experts*", + "model.layers.63.self_attn*", + "model.layers.64.mlp.shared_experts*", + "model.layers.64.self_attn*", + "model.layers.65.mlp.shared_experts*", + "model.layers.65.self_attn*", + "model.layers.66.mlp.shared_experts*", + "model.layers.66.self_attn*", + "model.layers.67.mlp.shared_experts*", + "model.layers.67.self_attn*", + "model.layers.68.mlp.shared_experts*", + "model.layers.68.self_attn*", + "model.layers.69.mlp.shared_experts*", + "model.layers.69.self_attn*", + "model.layers.7.mlp.shared_experts*", + "model.layers.7.self_attn*", + "model.layers.70.mlp.shared_experts*", + "model.layers.70.self_attn*", + "model.layers.71.mlp.shared_experts*", + "model.layers.71.self_attn*", + "model.layers.72.mlp.shared_experts*", + "model.layers.72.self_attn*", + "model.layers.73.mlp.shared_experts*", + "model.layers.73.self_attn*", + "model.layers.74.mlp.shared_experts*", + "model.layers.74.self_attn*", + "model.layers.75.mlp.shared_experts*", + "model.layers.75.self_attn*", + "model.layers.76.mlp.shared_experts*", + "model.layers.76.self_attn*", + "model.layers.77.mlp.shared_experts*", + "model.layers.77.self_attn*", + "model.layers.8.mlp.shared_experts*", + "model.layers.8.self_attn*", + "model.layers.9.mlp.shared_experts*", + "model.layers.9.self_attn*" + ] + } +} \ No newline at end of file diff --git a/model-00001-of-00045.safetensors b/model-00001-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..40ddfb0e94c02bc12bb5c4e438bdbb70a13dc22f --- /dev/null +++ b/model-00001-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b308a5f4e235ca979ed55a9d04246828d70bb5d1b7d8f6ae06fca76e9ff9852 +size 9994445096 diff --git a/model-00004-of-00045.safetensors b/model-00004-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3a5192c9aa8da9517dd6649b9588d452ef70675a --- /dev/null +++ b/model-00004-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01f3ab46fd77768ab9e1fb5e9fd07dc512c24d41b9f52407b27c12ce4ec216f0 +size 10000173744 diff --git a/model-00005-of-00045.safetensors b/model-00005-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5d0a0e73c3f5886fbe00f6659780ea9f3e0e6b6e --- /dev/null +++ b/model-00005-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9be72b5db9fb5e8c0bbae35e2ad33cc0371859523ff27d9cd1848e2b3a42ddf4 +size 9997417040 diff --git a/model-00007-of-00045.safetensors b/model-00007-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c2d1792e239dce399f6962e5c7bab84a12067ed7 --- /dev/null +++ b/model-00007-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:05a59d35a9d206e4c620ca39f0cb106d8e187538731082e1f0a6a8191a0ba6cd +size 10000178672 diff --git a/model-00008-of-00045.safetensors b/model-00008-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ff96a124648241a1fc0315f1b331b2ea12729136 --- /dev/null +++ b/model-00008-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6d4d6006655c3686be039267c7c2d4697352d84eca0e9a8d07eb42195a7d92cf +size 9997420960 diff --git a/model-00009-of-00045.safetensors b/model-00009-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..49eeaa3691294f530b4110217201a1983c2653c0 --- /dev/null +++ b/model-00009-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5774821daeb7430bbe2c2db1ad32a1ad7fd53443b0c4fea07e60fd6faa7ee513 +size 10000178680 diff --git a/model-00010-of-00045.safetensors b/model-00010-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a786997de265b359ec3e17a812787d1160065df1 --- /dev/null +++ b/model-00010-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a392b350bdb50a167d5ab53837aa479f1480b068942a1b285efd4b56fe079865 +size 10000178464 diff --git a/model-00011-of-00045.safetensors b/model-00011-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a1f2095c6d8070b5f7dadc644a5feec3f79f481c --- /dev/null +++ b/model-00011-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff5e4c6baf48da4eb731c49dd0094d38f5401452cedcc580019f5ef2af7276c5 +size 9992458640 diff --git a/model-00012-of-00045.safetensors b/model-00012-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fa95189c76d27b1fe69cc7dffc3473cbeed14232 --- /dev/null +++ b/model-00012-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ffc9d67ffc0c288b0f31c76edf36b3d3a011364b9c7367aa2673639d82e0d497 +size 9998062264 diff --git a/model-00013-of-00045.safetensors b/model-00013-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a9b17e42cc3adbdf4acc1645830a9b3ced9c20c6 --- /dev/null +++ b/model-00013-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c49d853912ad83cbc352ea263b4b78e721a4e23423116571dd4975149f2a13e +size 10000178448 diff --git a/model-00014-of-00045.safetensors b/model-00014-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cb6b223c9253139f416833b2572b92331daf666f --- /dev/null +++ b/model-00014-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eb566b59272d413a83cf54f839fea75a6fabafec5c8124bd6615ac3ce6061bdb +size 10000178872 diff --git a/model-00015-of-00045.safetensors b/model-00015-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2518890c54cc1f648a9eb3a54c0a35d6c0f2d6df --- /dev/null +++ b/model-00015-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e3db02c0d381364e408c94173c17bef531c81e19285835f1bf6d04b9d759032e +size 9997420752 diff --git a/model-00016-of-00045.safetensors b/model-00016-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..aeb4879635eea159b637f0383b19a12a11297661 --- /dev/null +++ b/model-00016-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:80745abab4ec4da3759762140be1b2ab5973f84ba9095ade02c7cf03f59ba4bd +size 10000178464 diff --git a/model-00017-of-00045.safetensors b/model-00017-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..4bc5cafc1c1642c98f6f9fc1436ec0d0ec9fe837 --- /dev/null +++ b/model-00017-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:92d0a2a583bdc1b602b26636ae0b7f48439653809aef633e890b9c1ea02b434c +size 10000178488 diff --git a/model-00018-of-00045.safetensors b/model-00018-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f887fa0da5b720a5c09519100575145974161107 --- /dev/null +++ b/model-00018-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dcaefca8fbb5670052e1854a2dd64d39cee860bfe0126e9a2fe7ffe53bc087aa +size 9997421136 diff --git a/model-00019-of-00045.safetensors b/model-00019-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ac94d1cf5d471ceffb686c309fdc468ade7bae8e --- /dev/null +++ b/model-00019-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:39c0cbde14edf84b41e66d7ad561cfe795b9176b15550c70aaa63e8371dc0fc5 +size 10000178848 diff --git a/model-00020-of-00045.safetensors b/model-00020-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..47afd4c9cdb1f2460402b15bd125a6f5ac96c1e9 --- /dev/null +++ b/model-00020-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aa1f07fe99f2851acbe1d5bf83d18ef2be1b022d4e9c8fda5aceae55ef91bc26 +size 10000178456 diff --git a/model-00021-of-00045.safetensors b/model-00021-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..188104430c0b7961f6eeb7c16f659311db9170b4 --- /dev/null +++ b/model-00021-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:045ee0c7947a31bacedec03d520ac5f42eb6500b739105d311ee6b1532ff5d44 +size 10000178984 diff --git a/model-00022-of-00045.safetensors b/model-00022-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..64903d7ac33369c78ec702bb1f23ae18432fa5ce --- /dev/null +++ b/model-00022-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8d269cf3234fa2e9f95127b10340380340c0724d6a5735e853f5624e9f2992f +size 9997420472 diff --git a/model-00023-of-00045.safetensors b/model-00023-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d71e6601f261081e0f85664fdea9a6ba04b5d269 --- /dev/null +++ b/model-00023-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d2067fa0c29a3882fc0b5b7ef7c74847dac0c55554529e9871377aa21d93c75e +size 10000178448 diff --git a/model-00024-of-00045.safetensors b/model-00024-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bb2c3b001b985cd9727011ed00cd77a5dd8b39cd --- /dev/null +++ b/model-00024-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c95f994e866a1638bbe5892db1c2774c5a8ae81e87f2b823c4a88144bd56061 +size 10000178704 diff --git a/model-00025-of-00045.safetensors b/model-00025-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f2be9318d4fc326fbebf172203a91438711eec00 --- /dev/null +++ b/model-00025-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11f203746c0f60e11f8bce237324aa94f063d3522fe6e4d3980a7b6465a96555 +size 9997420920 diff --git a/model-00026-of-00045.safetensors b/model-00026-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6409ff756312dbb8f15ef085d961ec5c13269234 --- /dev/null +++ b/model-00026-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:556f43ceb48e706f5f223ea2d0ce577880295cbcc0d4ee2f1ee98842bbdfe383 +size 10000178640 diff --git a/model-00027-of-00045.safetensors b/model-00027-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b62c794175afe6208ccec3b498f074b919cb62bc --- /dev/null +++ b/model-00027-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5af58bb2a9caec711592af4cc6c6315eb6118e9d0bfbd60dcd531b35ac1dd291 +size 10000178464 diff --git a/model-00028-of-00045.safetensors b/model-00028-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ab87c0aec75ac951712d6d031cc2d9cfad5c738c --- /dev/null +++ b/model-00028-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:634c6fd49ecc03d53b3d2b824c36d419bac4770f7f9809ff03cad1d4217a8e7b +size 9952615376 diff --git a/model-00029-of-00045.safetensors b/model-00029-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e0c4d38f2de1be5acd1e8eb26ac57fd81e061d88 --- /dev/null +++ b/model-00029-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4377e5c69de696407a5214d4729b5fe95ea0648c92072de87462d3c263c19091 +size 9995435208 diff --git a/model-00030-of-00045.safetensors b/model-00030-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9889e0adca53b8ef9cbae62c94f565633b7593d5 --- /dev/null +++ b/model-00030-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1aff393e8e22f76922de446fdd02c7b98b1d86bc3aa5ac9ac1a3f98dedefbf6b +size 10000178456 diff --git a/model-00031-of-00045.safetensors b/model-00031-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6852aa04e25cf1758318b08444396e39b89f2b5a --- /dev/null +++ b/model-00031-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:43da3079e18393d129e48b3010773aec8e66c13221429ce6fba3e95b2886134e +size 10000178912 diff --git a/model-00032-of-00045.safetensors b/model-00032-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..ff2e61615a2bbea4e9d16b9cf56bce4a7785150e --- /dev/null +++ b/model-00032-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:636e0d90e6bfbceabe4e0492367f2ac42051bb22c751b807ddbc89f012e2b5b6 +size 9997420696 diff --git a/model-00033-of-00045.safetensors b/model-00033-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..23fa8bcc1a1dcdb4c33247b79a1e04f4d0570e75 --- /dev/null +++ b/model-00033-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5088cacb3ce1ff16797a6a812d89c820b74741e4e7db2ab84bfedcc25bd7daf8 +size 10000178448 diff --git a/model-00034-of-00045.safetensors b/model-00034-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..5cf39edf21e0a1a8be9706da76393d18e309cbf7 --- /dev/null +++ b/model-00034-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5789392d340621cb0ffbccec89db8705e2566232b67994ea2d641b09b705265f +size 10000178552 diff --git a/model-00035-of-00045.safetensors b/model-00035-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2c64478c5ccf2aa2a8f60bec44122c570fa9e53e --- /dev/null +++ b/model-00035-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:703fd9a8c4ca7808d4a921e699918ee7e9582fcb566981462b94c5f3b3b26a6d +size 9997421080 diff --git a/model-00036-of-00045.safetensors b/model-00036-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8db528fc0ba99d7c2377d268eb89a1cca3b70b40 --- /dev/null +++ b/model-00036-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de2a6f8e65e9aea11ff429909f08e339e0810df7e5fd3856c202e3a9c8f5215f +size 10000178792 diff --git a/model-00037-of-00045.safetensors b/model-00037-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..12d692ea1f86b0f6e42d41a61494c4e31204864b --- /dev/null +++ b/model-00037-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7d825bb8a221be5282a582a25d7299b40cfcf5f2d898b43726840a164b059f19 +size 10000178456 diff --git a/model-00038-of-00045.safetensors b/model-00038-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b9aab63ed7d187efe207cd196e1e75534b070e2a --- /dev/null +++ b/model-00038-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:440475bcaec6a16a2fd2716726f5c5a4b857ada4ccd68b135c1d8b4d46054d5c +size 9995591328 diff --git a/model-00039-of-00045.safetensors b/model-00039-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bc6ac2fb5b587d7d16572dfa6979d51d0b8b9059 --- /dev/null +++ b/model-00039-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:332dd6c10fe55493c935dbfa7dfc754c67b4887428ad72958eba4e2af00a128f +size 9994929680 diff --git a/model-00040-of-00045.safetensors b/model-00040-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..bfccd96baab5c59c1a434a7fbc2a5c68fefe633e --- /dev/null +++ b/model-00040-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e47c772b688a3efa17a748931e742c69010c1155773714551159b061553019e9 +size 10000178456 diff --git a/model-00041-of-00045.safetensors b/model-00041-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7aebb116f59fa39890d998d4dbbed743b6a88d30 --- /dev/null +++ b/model-00041-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e2f65116d4949ef34c6bc106e0547e1e2719a8c65c5f04c472d9debcb692b2dc +size 10000178760 diff --git a/model-00042-of-00045.safetensors b/model-00042-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..8be6458f9a8aad0ca7ca8c13da125986a39b09bc --- /dev/null +++ b/model-00042-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:65a54d32bfa45e54588d74a027118a68ff5c522006b3b51b2b6dc1a3eeeb9995 +size 9997420856 diff --git a/model-00043-of-00045.safetensors b/model-00043-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..535a1db14027f23f8c44a487cc3dc73454d95dd5 --- /dev/null +++ b/model-00043-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01c5d275ae93da06ee97a28ab91fe245193ceadcc397795c6df831917c493e03 +size 10000178584 diff --git a/model-00044-of-00045.safetensors b/model-00044-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b7def6b450869e566f66444b0d865aa45a1f20ca --- /dev/null +++ b/model-00044-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c2d9f25f86d7b7dddae264bad40327ef70365c8f4e2581101b816322f7e9ac0 +size 10000178456 diff --git a/model-00045-of-00045.safetensors b/model-00045-of-00045.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d92ee6c3843893dabf3a96aecbd060b96a4f50a1 --- /dev/null +++ b/model-00045-of-00045.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c887d5e75a5e4fcba9e4a1ed2310e4ec599d318783a53212b5d99a2a5738bc8 +size 6080226488 diff --git a/model.safetensors.index.json b/model.safetensors.index.json new file mode 100644 index 0000000000000000000000000000000000000000..5aabd60107df9883856aeaf90598e4d3229d22ed --- /dev/null +++ b/model.safetensors.index.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59bc81fea793850940f43c3ee89aa883a3a71adbfb3d19c97ac7046f0733c5da +size 22148477 diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..aba40197a4cdb5607f4ab7a05fb0a4ee8054fd6d --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19e773648cb4e65de8660ea6365e10acca112d42a854923df93db4a6f333a82d +size 20217442 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..affcef6fd609d9a13657a09ece48bfe40ee415fa --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,34 @@ +{ + "backend": "tokenizers", + "clean_up_tokenization_spaces": false, + "do_lower_case": false, + "eos_token": "<|endoftext|>", + "extra_special_tokens": [ + "<|endoftext|>", + "[MASK]", + "[gMASK]", + "[sMASK]", + "", + "", + "<|system|>", + "<|user|>", + "<|assistant|>", + "<|observation|>", + "<|begin_of_image|>", + "<|end_of_image|>", + "<|begin_of_video|>", + "<|end_of_video|>", + "<|begin_of_audio|>", + "<|end_of_audio|>", + "<|begin_of_transcription|>", + "<|end_of_transcription|>" + ], + "is_local": true, + "local_files_only": false, + "model_max_length": 202752, + "model_specific_special_tokens": {}, + "pad_token": "<|endoftext|>", + "padding_side": "left", + "remove_space": false, + "tokenizer_class": "TokenizersBackend" +}